The world of advertising is awash in data. Too much data. In this article, we trim it down. We separate the signal from the noise. We focus on the most important metrics, and we share a simple framework that optimizes for understanding the big picture, while being able to quickly spot opportunities for improvement.
Let’s start with the most fundamental equation of ad ops. In physics, Newton’s famous postulate tells us that Mass x Acceleration = Force. In ad ops, the governing law of the universe is:
(Pageviews x CPM per Pageview)/1000 = Revenue
The equation elegantly ecompasses the most critical aspects of a website: traffic and monetization. With these numbers, you can see a high-level view of performance.
There are alternate ways to slice this pie. Some publishers focus on users and revenue per user. Revenue per user is a fine metric, but it severely blurs the line between traffic and monetization. Other publishers put the emphasis on CPM per Ad Impression as the top level metric, rather than CPM per Pageview. The challenge here is that key shifts in the industry, like the move from desktop to mobile or the rise of adblock, can be overlooked if the ads per page is not factored into the equation. Additionally, the performance of individual units can distract from the overall goal of revenue optimization.
In our standard daily, weekly, and monthly reporting, we segment these top-level numbers by geography and platform. We find it helpful to see performance on desktop, mobile, and apps separated, and performance in the US distinguished from international.
Finally, we aim to add context to the numbers when we report on them, so that they can be interpreted. We always include the percent change from the same period last year, month, or week, and if we have a specific target, the percent of our goal. Below are some sample reports:
The equation above can be broken down into smaller components. Pageviews, for example, can be split into its own equation:
Users x Pageviews per User = Pageviews
and CPM per Pageview can be broken down into:
CPM per Ad Impression x Ads per Page = CPM per Pageview
Putting this all together, we have a simple framework for understanding traffic and monetization, which we can visualize as follows:
These metrics correspond well to real-world questions every publisher must ask themselves:
Of course the precise metrics publishers track will vary. Regardless of the approach, however, the metrics should:
Digging into these metrics is invaluable for planning, diagnosing problems, and spotting opportunities. Let’s imagine we notice a change in the CPM per Pageview yesterday as compared to the same day in the prior week and want to understand what happened. Using the framework, we’d run a report to analyze where the decline occurred, in the CPM per Ad Impression or the Ads per Page. Here's an example analysis:
Ok, in this instance, the CPM per Ad Impression declined. Having identified the source of the change, we’d run segmentation reports by partner, ad unit, and geography to better understand why. Perhaps a partner lowered their CPM? Or maybe a particular geography took a hit? Here’s an example of how the analysis might unfold:
Above we see that the revenue from Partner 5 plunged. To find out why, We might look at whether the drop occurred because of a slide in a particular ad unit:
Looks like the decline occurred across all ad units. So perhaps we check to see if the CPM decrease was isolated to a particular geography:
In this analysis, we notice that the drop occurred in English-speaking international countries. At this point, we’d probably check a number of possible explanations for the isolated drop in this region: tag errors, campaigns ending, fill rates, click-though-rates, and other factors that influence performance.
The key here is that we always start with the high level, most important trend--like a drop in CPM per pageview--and then peel away the layers of possible reasons as we dig deeper. Starting with the top-level, high-value fluctuations prevents the down-the-rabbit-hole scenario that can arise when focusing less-critical metrics.
Deep-dives should also be conducted during times of planning. Analyzing the drivers of revenue, and doing it by segment, can lead to insights about where there are opportunities for improvement. Maybe the ads per page is too low on mobile. Perhaps the CPM per Ad Impression declined internationally. These types of questions can easily be answered when doing a deep dive.
Do you prioritize different metrics or have a different way of looking at performance? Share what you focus on in the comments section below.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form
Six months ago we launched an A/B testing framework for our site, SpanishDict.com, that has allowed us to make ad ops decisions smarter, faster, and with measurable improvements to revenue. Our fir...
In the past two months, we've radically overhauled our programmatic advertising stack on SpanishDict.com, shifting our focus to header bidding advertisers--advertisers that submit the price they wi...